Research On Key Technology Of Machine Vision For Surface Paint Machine Control System

Surface painting is a process of applying paint,ink,and paint on the surface of metal products and is used in many places in life.In response to this demand,various types of equipment such as manual paint,semi-automatic spot coater,and automatic spot coater have been gradually developed,but these devices have low adaptability to workpieces,high environmental requirements,and high system complexity.Considering the increasing market demand for spot paint,the development of a fully-automated visual spot paint system that is easy to operate and has excellent performance is imminent.This paper aims to develop a fully-automated machine vision system for equipment,focusing on the realization of functional modules with high environmental applicability,high stability,and ease of use.The main work of the paper is as follows:1)Based on the application requirements,design indicators for visual point-painting equipment are proposed,and hardware plans for the system are planned,including vision systems,motion control systems,and mechanical systems.The camera,lens,and light source were selected and analyzed for the components of the vision system.2)This paper researched image calibration techniques suitable for spot painting processes.The commonly used calibration methods generally require the production of a high-precision calibration board,and the calibration process is complicated and the procedures are cumbersome.The current mainstream calibration method was discussed.In the actual environment of the visual spot painting system,an improved simplified calibration method was proposed in combination with the Tsai two-step method.The feasibility of the simplified calibration method was verified by calibration experiments.3)The image matching algorithm applied to the surface paint process was studied.For the visual surface paint application,the workpiece is complex in appearance,rich in features,partially with spatial curved surfaces or reflective surfaces,and requires high matching algorithm.Common template matching methods and related edge detection algorithms are studied.Based on the performance comparison between gray-scale and shape-based template matching,a shape template matching method suitable for spot-paint applications was developed,and a method for increasing the matching speed was realized.4)According to the application requirements,the software framework of the system was designed and divided into modules,and the system’s interactive interface was set up under Microsoft Visual Studio 2013.The focus is on the programming process with the highest time-consuming teaching process in point-paint applications,which increases the speed of the teaching process and makes maintenance easier.5)Through experiments,the actual performance of the functional blocks in this paper was tested and analyzed.Compared with the existing systems,the verification shows that the research and development system has obvious advantages.